Improving drug selectivity for target tissue is an established goal in the medical arts. In general, it is desirable to deliver a drug selectively to its target, so that dosage and, consequently, side effects can be reduced. This is particularly the case for toxic agents such as anti-cancer agents because achieving therapeutic doses effective for treating the cancer is often limited by the toxic side effects of the anti-cancer agent on normal, healthy tissue. The problems relating to lack of drug selectivity can be exemplified by Taxol.
Taxol® (paclitaxel) was first isolated in 1971 from the bark of Taxus brevifolia and was approved in 1992 by the US Food and Drug Administration for treatment of metastatic ovarian cancer and later for breast cancer. Its mechanism of action is believed to involve promoting formation and hvperstabilization of microtubules, thereby preventing the disassembly of microtubules necessary for completion of cell division. It also has been reported that Taxol induces expression of cytokines, affects the activity of kinases and blocks processes essential for metastasis, in as yet uncharacterized mechanisms of action.
Taxol has attracted unusually strong scientific attention, not only because of its unique antiproliferative mechanism of action, but also because it is active against nearly all cancers against which it has been tested and because it has been discovered to be an analog of numerous closely related compounds occurring naturally. These compounds, taxanes, are now recognized as a new class of anticancer compounds.
Taxol's strength against cancers of diverse tissue origin also represents a significant drawback. An ideal anticancer agent has tissue specificity, thereby reducing side-effects on normal (dividing) cells. Taxol analogs with tissue specificity therefore are desired. Another drawback of Taxol is its extreme insolubility. Taxol can be administered effectively in a solvent including Cremophor® EL (polyoxyethylated castor oil), which combination can provoke severe hypersensitive immune responses. As a result of these drawbacks, and also as a result of the potential for modifying Taxol at numerous sites as demonstrated by other naturally-occurring taxanes with anticancer activity, a search for more selective taxanes was launched.
To date, more than 200 taxanes have been synthesized (or isolated) and tested in vitro or in vivo for anticancer activity. The results, however, have been so disappointing that the National Cancer Institute (NCI) generally no longer is interested in testing Taxol analogs. In general with Taxol analogs, the solubility problems remain, and/or potency is sharply reduced, and/or selectivity is not improved, and/or the ratio of the median toxic dose to the median effective dose (“therapeutic index”) is unacceptably reduced.
Taxol has the following formula: 
Taxanes have the basic three ring structure (A, B and C), substituted or unsubstituted.
Taxol's carbons are numbered conventionally as follows: 
Based upon the taxanes tested to date, as many questions have been raised as have been answered, and general rules have not been fashioned easily in predicting selectivity, activity and solubility. Firstly, no rules have emerged regarding selectivity. Those taxanes that are strongly active appear to have activity as broad as Taxol's activity, and no headway appears to have been made in terms of developing a more selective Taxol analog.
Some information about activity has emerged. Numerous substitutions have been made at C7, C9, C10, C19, R1 and combinations thereof while retaining significant, but usually reduced, activity. Substitutions at C2, C4 and 2′OH, however, are generally not tolerated. These conclusions are only generalities, for example, because some substitutions at C9-C10 (cyclic derivatives) are not tolerated and some substitutions at C2 (meta substitutions on the phenyl) are tolerated. Likewise, the C13 side chain and, in particular, the 2′OH are required, although the minimum structural requirements of the side chain have not been determined for therapeutic efficacy.
Attempts to improve Taxol's solubility have not resulted in successful clinical products. One approach has been to manufacture prodrugs of Taxol, which prodrugs undergo in vivo transformation into Taxol and some other product. Attempts were made to esterify the C7 hydroxy and 2′ hydroxy groups, with the hope that the bond would be stable in solution (to permit preferred administration modes—i.v. over at least 24 hours) but would cleave readily in vivo. The groups tested were all hydrophilic and included amines, short carboxylic acids (using e.g. succinic anhydride and glutaric anhydride), sulfonic acids, amino acids and phosphates. Generally, activity was reduced although some success was obtained with certain derivatives. Again, no particular pattern emerged permitting one to predict reliably which groups could be substituted on Taxol to yield a therapeutically useful product, although it was suggested that the 2′ OH derivatives may cleave more easily than the C7 OH derivatives.
Several other factors add to the problem of predicting which Taxol analogs will be effective. Multiple mechanisms of action have been proposed in the literature, and a change in one position may have no effect on activity on one such mechanism but may eliminate activity on another mechanism. In addition, changes that favorably influence activity may unfavorably influence bioavailability. For example, Taxol affects microtubule formation inside a cell, but a change in structure that increases intracellular activity may adversely affect the ability of Taxol to gain entry into a cell. Taxol also is known to bind to proteins, and the effect on activity that results from a change in Taxol's binding to protein (in terms of conformation, cellular absorption and solubility) is unknown.
It has been reported that Taxol does not get into the brain, apparently excluded by the blood brain barrier. It is not known why this is so, as Taxol is lipophilic, gets into cells and might be expected to cross the blood brain barrier.
Among the most promising of the two hundred analogs tested is Taxotere® (docetaxel), because of its slightly increased activity and solubility. Oddly, however, Taxotere differs from Taxol at sites which typically do not have a strong influence on activity, and one would not predict the improvements in Taxotere from these differences, even in hindsight.
Taxotere has the following formula: 
Fatty acids previously have been conjugated with drugs to help the drugs as conjugates cross the blood brain barrier. DHA (docosahexaenoic acid) is a 22 carbon naturally-occurring, unbranched fatty acid that previously has been shown to be unusually effective, when conjugated to a drug, in crossing the blood brain barrier. DHA is attached via the acid group to hydrophilic drugs and renders these drugs more hydrophobic (lipophilic). DHA is an important constituent of the brain and recently has been approved as an additive to infant formula. It is present in the milk of lactating women. The mechanism of action by which DHA helps drugs conjugated to it cross the blood brain barrier is unknown.
Another example of the conjugation of fatty acids to a drug is the attachment of pipotiazine to stearic acid, palmitic acid, enanthic acid, undecylenic acid or 2,2-dimethyl-palmitic acid. Pipotiazine is a drug that acts within the central nervous system. The purpose of conjugating pipotiazine to the fatty acids was to create an oily solution of the drug as a liquid implant for slow release of the drug when injected intramuscularly. The release of the drug appeared to depend on the particular fatty acid selected, and the drug was tested for its activity in the central nervous system.
Lipidic molecules, including the fatty acids, also have been conjugated with drugs to render the conjugates more lipophilic than the drug. In general, increased lipophilicity has been suggested as a mechanism for enhancing intestinal uptake of drugs into the lymphatic system, thereby enhancing the entry of the conjugate into the brain and also thereby avoiding first-pass metabolism of the conjugate in the liver. The type of lipidic molecules employed have included phospholipids, non-naturally occurring branched and unbranched fatty acids, and naturally occurring branched and unbranched fatty acids ranging from as few as 4 carbon atoms to more than 30 carbon atoms. In one instance, enhanced receptor binding activity was observed (for an adenosine receptor agonist), and it was postulated that the pendant lipid molecule interacted with the phospholipid membrane to act as a distal anchor for the receptor ligand in the membrane micro environment of the receptor. This increase in potency, however, was not observed when the same lipid derivatives of adenosine receptor antagonists were used, and generalizations thus were not made possible by those studies.